A seminar by Marta Revuelta

Saturday 5th October 2019 at 3 pm


Facial recognition cameras scan for distinct and specific features, such as facial shape, to create a detailed biometric map, a picture from these cameras is like a fingerprint, but taken without knowledge or consent. Facial recognition technology uses cameras and software to associate live images of people in public with images on a watch list. The images in a watch list can come from a variety of sources and may not include only faces of suspected criminals. As debates and demonstrations against the controversial use of facial recognition technologies are proliferating around the world, police and government forces are increasingly using facial analysis technology to monitor citizen pacific protests and riots. Recent protests in Hong Kong have exposed this surveillance infrastructure based on this technology put in place by governments, highlighting their potential misuse scenario applied as part of a mass citizen surveillance.


In parallel, a facial recognition research funded by the Office of the Director of National Intelligence (ODNI) and Intelligence Advanced Research Projects Activity (IARPA) provides a method capable of producing detailed 3D reconstructions of the face from photos taken in conditions of extreme vision. This method then consists of estimating the plausible details of the face in places where it is hidden in order to reconstruct it in its entirety. It is an advance that in the near future could defeat attempts to camouflage anonymity protesters and activists.


The 3D Facial Weaponization seminar consists in creating an experimental space to investigate the tense relationship between demonstrators and facial recognition. It also provides an opportunity to review contemporary research and test new algorithms in the field of facial reconstruction, with the objective of maintaining a dialogue on freedom to demonstrate and the use of AI in mass surveillance.